BI 10773 add-on to Metformin in Patients With Type 2 Diabetes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for BI 10773 add-on to Metformin in Patients With Type 2 Diabetes (NCT00749190), a registered phase 2 study. The primary outcome is Change From Baseline in HbA1c After 12 Weeks of Treatment.

2. Clinical Question

The study was designed to investigate efficacy, safety, and pharmacokinetics of BI 10773 when added to ongoing metformin in patients with type 2 diabetes and insufficient glycemic control. The registry also describes sitagliptin as an open-label add-on treatment arm. The primary endpoint was change from baseline in HbA1c after 12 weeks.

This question has two parts that should be kept distinct. The endpoint asks about change in a continuous measure, HbA1c, while the study design sets out several treatment groups for comparison. The record indicates that the BI 10773 and placebo group means were adjusted and derived from a model excluding the open-label sitagliptin arm. The sitagliptin mean came from a model including that group and placebo. Details of the models are not reported in the registry record.

The registry record does not include the numerical primary outcome estimates, their uncertainty, or the statistical comparisons. It therefore does not establish from the posted results how the groups differed or how precisely any difference was estimated.

3. Trial Design

This was a phase II, randomized, parallel-group study with an overall enrollment of 495. It concerned patients with type 2 diabetes and insufficient glycemic control despite metformin therapy. The planned treatment period for the primary endpoint was 12 weeks. The record describes five once-daily oral BI 10773 dose groups, a placebo group, and an additional open-label sitagliptin group. The BI 10773 and placebo groups were described as double blind; sitagliptin was open label.

The registry lists BI 10773, placebo, and sitagliptin as interventions. In the adverse-event group listings, the BI 10773 dose groups are named Empagliflozin groups. The primary endpoint description also refers to empagliflozin arms. The record does not explain the naming relationship between BI 10773 and empagliflozin, so the naming is reported as it appears in the registry.

Although the study is described as randomized, the method of allocation is not reported in the registry record. The record does not report how the random sequence was generated, how allocation was concealed, or whether the groups had equal planned allocation. The blinding description applies to the BI 10773 and placebo comparison, while the sitagliptin arm was open label. Further details about who was blinded are not reported.

The primary endpoint was change from baseline in HbA1c after 12 weeks. Listed secondary outcomes include fasting plasma glucose, HbA1c over time, attainment of specified HbA1c targets or reductions, fasting plasma insulin, HOMA-IR, HOMA-%B, and body weight. The study objective also mentions pharmacokinetics, but detailed pharmacokinetic results are not reported in the registry record.

The study is marked completed. Its start and primary completion are recorded as 2008-08 and 2009-10, respectively. These design details describe the setting in which the outcomes were intended to be assessed; they do not substitute for the missing numerical results.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Only the first 6 of 7 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full analysis set (FAS) consisting of all randomized patients who were treated with at least one dose of study drug and has a baseline measurement of the primary endpoint. The imputation method used was a modified last observation carried forward (LOCF) approach which used linear interpolation, LOCF and worst observation carried forward (WOCF).

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Treated Set (TS) consisting of all patients who were dispensed study medication and were documented to have taken at least one dose of investigational treatment.

Only the first 6 of 8 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in HbA1c After 12 Weeks of Treatment, measured over Baseline and 12 weeks.

The primary endpoint is change from baseline in HbA1c after 12 weeks of treatment, expressed as a percentage of HbA1c. The registry describes the measure as a mean and identifies its dispersion as the standard error. It also states that adjusted means are displayed, although the numerical values are not reported in the registry record.

HbA1c is used as a measure of glycemic control over time. As a continuous endpoint, it can describe the average level of change across a treatment group rather than reducing each participant’s result to a yes-or-no category. A change-from-baseline measure relates the follow-up value to each participant’s starting value, making the direction and magnitude of change the focus of the endpoint.

That endpoint is distinct from the listed secondary outcomes, which include fasting plasma glucose and the proportions of patients reaching specified HbA1c categories. Continuous and categorical outcomes answer different questions: a group mean describes average change, while a proportion describes how many participants meet a defined threshold. The registry does not report the primary endpoint estimates, so the magnitude and precision of the group changes cannot be assessed from the record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes a continuous primary endpoint and reports that adjusted means were used. In general, a mean summarizes the central value of a continuous outcome in a group. An adjusted mean is an estimate from a statistical model that accounts for selected variables or structure. Adjustment can make a comparison conditional on the model’s terms, but the meaning depends on which terms were included and how the model was specified. Those details are not reported in the registry record.

The record distinguishes the models used for the treatment groups. The placebo and BI 10773 means were from a model excluding the sitagliptin open-label arm. The sitagliptin mean was from a model containing sitagliptin and placebo. This describes how the means were derived, but does not identify the model’s covariates, estimation method, or exact contrast. The numerical estimates and comparative tests are not reported in the registry record.

The stated dispersion measure is the standard error. In general, a standard error describes how much an estimated quantity, such as a group mean, would vary across repeated samples under the model. It is not the same as the standard deviation, which describes variation among individual observations. A confidence interval can be constructed from an estimate and its uncertainty under specified assumptions; the registry record does not report primary endpoint confidence intervals.

What does change from baseline represent?

For a change-from-baseline endpoint, each participant’s follow-up measurement is interpreted relative to that participant’s starting measurement, and the group summary describes the resulting changes. This can focus an analysis on improvement or worsening relative to baseline rather than on the follow-up value alone. The record specifies baseline and 12 weeks as the primary endpoint timeframe. It does not report the individual-level data or the numerical group summaries needed to assess the distribution of changes.

What does the analysis population mean?

The primary endpoint population is described as the full analysis set. It includes randomized patients who were treated with at least one dose of study drug and had a baseline measurement of the primary endpoint. This definition excludes randomized participants who do not meet those conditions, but the number excluded and the resulting group sizes are not reported in the registry record. The label alone does not establish how missing follow-up measurements were handled beyond the stated imputation approach.

How should the imputation description be read?

The registry says a modified last observation carried forward approach was used, involving linear interpolation and LOCF. In general, imputation substitutes an estimated value for a missing measurement so that an analysis can include participants without a complete observed endpoint. Last observation carried forward uses an earlier observed value as a substitute for a later missing value. Linear interpolation estimates a value between observations. The registry calls its approach modified and mentions both methods, but does not report the exact algorithm, the circumstances in which each method was applied, or sensitivity analyses using alternative assumptions.

Imputation requires assumptions about unobserved outcomes. If a participant’s missing value differs systematically from the substituted value, the resulting estimate may not reflect the unobserved outcome distribution. This is a general limitation of methods that fill in missing measurements; it does not show that the estimates in this study were biased in a particular direction. The registry record does not provide enough detail to assess the assumptions or their impact.

What can be inferred from the design description?

Randomization is stated, and the study is described as parallel group. In general, random allocation can help balance measured and unmeasured baseline characteristics across groups, while a parallel design assigns participants to a treatment group for the comparison period. The allocation procedure and concealment methods are not reported in the registry record, so those implementation details cannot be evaluated. The sitagliptin arm was open label, whereas BI 10773 and placebo were described as double blind.

The record does not report a statistical analysis plan, the model’s detailed assumptions, handling of multiplicity, or interim analyses. No sample-size justification is reported in the registry record. The overall enrollment figure does not by itself establish the planned power for the primary comparison or the number analyzed in each group.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Placebo", "Empagliflozin 1 mg", "Empagliflozin 5 mg", "Empagliflozin 10 mg", "Empagliflozin 25 mg", "Empagliflozin 50 mg")
n   <- c(71, 71, 71, 71, 70, 70)
m   <- c(0.15, -0.09, -0.23, -0.56, -0.55, -0.49)
sp  <- c(0.08, 0.08, 0.08, 0.08, 0.08, 0.08)
sd  <- sp * sqrt(n)      # standard errors converted to SDs
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Placebo  71  0.150  0.674  -0.010  0.310
CARM  Empagliflozin 1 mg  71  -0.090  0.674  -0.250  0.070
CARM  Empagliflozin 5 mg  71  -0.230  0.674  -0.390  -0.070
CARM  Empagliflozin 10 mg  71  -0.560  0.674  -0.720  -0.400
CARM  Empagliflozin 25 mg  70  -0.550  0.669  -0.710  -0.390
CARM  Empagliflozin 50 mg  70  -0.490  0.669  -0.650  -0.330

11. Multiplicity, Missing Data and Analysis Populations

The study includes several BI 10773 doses, a placebo group, an open-label sitagliptin group, one primary endpoint, and multiple listed secondary outcomes. In general, testing many outcomes or many treatment contrasts creates a multiplicity issue: the chance of at least one apparently unusual result can increase when numerous tests are considered. A prespecified procedure can address this, but the registry record does not report a multiplicity adjustment, a hierarchy for testing outcomes, or which comparisons were confirmatory.

The registry identifies change in HbA1c after 12 weeks as the primary endpoint and lists secondary outcomes, but the statistical testing plan is not reported in the registry record. It is therefore not possible to determine which secondary findings, if any, were adjusted for multiple testing or how results across the dose groups were intended to be interpreted.

For missing data, the record describes a modified LOCF approach involving linear interpolation and LOCF. The exact implementation, amount and pattern of missingness, and any sensitivity analyses are not reported in the registry record. These details matter because imputation relies on assumptions about outcomes that were not observed.

The primary endpoint population is defined as the full analysis set: randomized patients treated with at least one dose and with a baseline primary endpoint measurement. The number meeting that definition, the numbers excluded, and the numbers contributing to each endpoint are not reported in the registry record. The record also does not report a separate per-protocol population or how protocol deviations were handled.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: coronary artery disease (1.4%), chest pain (1.4%), cholecystitis acute (1.4%), pneumonia (1.4%), pyelonephritis (1.4%).

The registry lists serious adverse events for the placebo group and each treatment group. For placebo, 2 of 71 participants at risk were affected by a serious adverse event. The corresponding counts were 0 of 71 for Empagliflozin 1 mg, 3 of 71 for Empagliflozin 5 mg, 1 of 71 for Empagliflozin 10 mg, 2 of 70 for Empagliflozin 25 mg, 3 of 70 for Empagliflozin 50 mg, and 0 of 71 for the open-label sitagliptin group. The record also lists other adverse events separately; those figures are not serious-event counts.

These are descriptive counts within the registry’s listed groups. They do not establish that an event was caused by treatment, and they should not be treated as adjusted comparisons. The record does not report definitions or adjudication details for these categories, exposure-adjusted rates, or statistical tests for between-group safety differences. Death counts are not reported in the registry record.

The listed serious-event terms include coronary artery disease, chest pain, acute cholecystitis, pneumonia, pyelonephritis, and arthropod bite, each recorded as 1.4%. The registry record does not provide enough context to infer comparative risk from these term listings. Counts and terms describe the posted safety information; they do not alone establish a treatment effect.

13. What the Result Does, and Does Not, Mean

This record documents a phase II randomized parallel-group study of BI 10773 added to ongoing metformin, with placebo and open-label sitagliptin groups. It identifies change from baseline in HbA1c after 12 weeks as the primary endpoint and describes the analysis population and imputation approach. These features clarify what the study was intended to assess and how the primary endpoint was summarized at a high level.

The numerical primary endpoint results are not reported in the registry record. Without group estimates, uncertainty measures, and comparative test results, it is not possible to quantify the observed changes or assess their precision from the posted information. The record therefore does not permit a conclusion about the size or direction of any difference between treatment groups.

The listed secondary outcomes and safety counts add context about what was measured, but they do not replace the missing primary endpoint results. The presence of an outcome in the registry is not evidence that it favored any group. Likewise, descriptive adverse-event counts do not establish causation or a comparative safety effect.

Statistically, the distinction is between a study’s design and its results. Randomization and a prespecified endpoint can define a comparison, but an interpretation of the result requires the estimates and information about their uncertainty. The registry record describes some analysis features, including adjusted means and a modified LOCF approach, while leaving other details—such as model specification and multiplicity procedures—not reported in the registry record.

The defensible reading is therefore narrow: the study was designed to evaluate HbA1c change and other outcomes over 12 weeks in patients receiving metformin, and the record contains selected safety counts. The posted information does not establish the magnitude, precision, or statistical significance of the primary endpoint comparisons.

14. Why This Trial Matters Statistically

This study is statistically notable for its multi-group structure: five BI 10773 dose groups, a placebo group, and an open-label sitagliptin arm. A design with several groups can address a range of treatment questions, but it also makes clarity about the planned comparisons important. The registry record states that the sitagliptin mean came from a different model set than the placebo and BI 10773 means, which is relevant when interpreting estimates side by side.

The primary endpoint is a continuous change-from-baseline measure, and the record reports adjusted means with standard errors as the intended summary. That combination can communicate both an estimated group-level change and its uncertainty. In this record, however, the numerical estimates and uncertainty measures are not reported, limiting statistical interpretation despite the endpoint and analysis population being described.

The record also illustrates why analysis details matter. It names a full analysis set and a modified LOCF approach involving linear interpolation and LOCF, but does not report implementation details, missingness patterns, or sensitivity analyses. Similarly, multiplicity handling and model specification are not reported. Those omissions constrain evaluation of the robustness and interpretation of any estimates that might otherwise be available.

Finally, the adverse-event listings show why descriptive counts must be distinguished from comparative inference. Group counts can describe events recorded in each arm, but without additional analytic context they do not establish causation or an adjusted difference. The trial is therefore informative about the intended statistical questions and some analysis choices, while the numerical efficacy results needed to assess those questions are not reported in the registry record.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The principal limitation for interpreting the primary endpoint is that its numerical results are not reported in the registry record. The adjusted mean changes, standard errors, confidence intervals, group-specific analysis counts, and comparative tests are absent from the record used for this summary. Consequently, the magnitude and precision of the primary comparison cannot be assessed.

Important analytical details are also not reported in the registry record. These include the statistical model specification, covariates, exact contrasts, allocation procedure, and handling of multiplicity. The record describes the use of adjusted means, but does not provide enough detail to evaluate the assumptions or reproduce the analysis.

The missing-data description is limited to a modified LOCF approach involving linear interpolation and LOCF. The degree and pattern of missingness, the exact imputation rules, and sensitivity analyses are not reported. The full analysis set is defined, but its size and the number of participants contributing to each outcome are not reported.

The sitagliptin arm was open label, and its mean was derived from a model containing sitagliptin and placebo, unlike the model described for placebo and BI 10773. The record does not report details that would permit a full assessment of comparability between these modeling sets. Safety listings are descriptive, and the record does not report comparative safety analyses, exposure-adjusted rates, or death counts.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: BI 10773 add-on to Metformin in Patients With Type 2 Diabetes — NCT00749190.
• Registry record last updated 2014-06-13; status last verified 2014-05.
• Results first posted 2014-06-13.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

NCT00749190 is a completed phase II randomized parallel-group study with an overall enrollment of 495. It enrolled patients with type 2 diabetes and insufficient glycemic control despite ongoing metformin therapy. The study evaluated five once-daily oral BI 10773 doses against placebo over 12 weeks and included an additional open-label sitagliptin arm.

The primary endpoint was change from baseline in HbA1c after 12 weeks. Adjusted means were described, with distinct model sets for the placebo and BI 10773 groups versus sitagliptin and placebo. The primary endpoint population was the full analysis set, and the record describes a modified LOCF approach involving linear interpolation and LOCF.

Numerical primary endpoint results, model details, multiplicity procedures, missingness summaries, and death counts are not reported in the registry record. Serious adverse-event counts are listed by group. The record therefore documents the study design and selected methods and safety information, but does not provide the numerical efficacy estimates needed to assess the primary comparison.